Differences in Risk Factors for Graft and Patient Survival Among Renal Transplant Recipients with Type 2 Diabetes Mellitus. Generation of a T2DM Specific Risk Calculator. (July 2018)

• Record Type: Journal Article
• Authors: Goodman, David J.; Ullah, Shahid; McDonald, Stephen P
• Is Part Of: Transplantation. Volume 102(2018)Supplement 7S-1
• Publication Date: 2018-07
• Subjects: Transplantation of organs, tissues, etc -- Periodicals; Transplantation immunology -- Periodicals; 617.95
• DOI: 10.1097/01.tp.0000542668.89146.f1
• Languages: English
• ISSNs: 0041-1337

Abstract

Abstract : Introduction: While the risk factors that influence graft and patient survival (G/PS) in non-diabetic recipients has been clearly defined it is not clear if the same factors are relevant in recipients with type 2 diabetes mellitus (T2DM). Materials and Methods: Data for all adult first transplant recipients from 2005-2014 was extracted from ANZDATA Registry. Age, gender, BMI, smoking, history of coronary artery disease (CAD), cerebrovascular disease (CVD) and peripheral vascular disease (PVD), indigenous status, dialysis duration, donor source and HLA matching were included. A Cox proportional hazard model was used to predict G/PS. Results: Of the 7010 patients transplanted 15% had T2DM. T2DM recipients were older, more obese (BMI>30), have more CAD, CVD and PVD. T2DM had longer dialysis time pre-transplant, and received a higher proportion of deceased donor kidneys. More Indigenous Australians were in T2DM group. (All comparisons significant P<0.001) Both G/PS improved in T2DM when 2010-2014 were compared with 2005-2009 outcomes (P=0.04 HR 0.7 & P=0.03 HR 0.62) but no improvement was observed in Non-T2DM, G/PS, (P=0.66 HR 0.94 & P=0.91 HR 0.94). Of the medical co-morbidities, only PVD was a significant risk factor in T2DM (G/PS P=0.03 P=0.01) whereas smoking, CAD & PVD were risk factors for non-T2DM G/PS. There was an interaction between CAD and T2DM for both G/PS and recipient age and T2DM for PS. We then developed 3 models to predict G/PS of increasing complexity. Model 1 (4 variables) included recipient age and gender, donor age and indigenous status. Model 2 (9 variables) added smoking, CAD, CVD, PVD and biopsy proven diabetic nephropathy. Model 3 (13 variables) added donor source, dialysis time, cold ischaemia time and HLA mismatch. The inclusion of more variables improved the predictive ability for 5y G/PS from C statistics of 0.61 for model 1 to 0.66 for model 3. Discussion: In a large cohort of T2DM recipients, outcomes have improved but remain inferior to non-T2DM. The demographics differ significantly between the two groups and conventional risk factors for non-T2DM cannot explain the difference in G/PS. Risk factors analysis was different in T2DM recipients. Conclusions: The risk factors that influence G/PS vary between T2DM and non-T2DM recipients. We have generated a simple T2DM specific tool with good predictive ability to assist clinicians in determining suitability for transplantation. The models with variables apparent only at the time of transplant listing (model 1 & 2) have limited predictive ability. We plan to validate model 3 using data from international data registries and produce an "app" to make the model user friendly.